7 Mar, 10 | by Bob Phillips
I’m fairly sure you’ll remember the RAMbo method of reviewing the validity of single randomised controlled trials. And so I think that many readers will have been having sleepless afternoons, struggling through the lengths of a ‘User’s Guide’ checklist for systematic reviews thinking “Which action hero can rescue me from this mire?”.
Or perhaps not.
But whichever, there is another rapid review acronym you should all learn to do it: FAST.Which is, in itself, rather nice, isn’t it?
F – Find all studies? (Did they – if they didn’t, and you can think of where they might well have lost stuff, then you can throw it out.)
A – Appraise them right? (If they were looking at the effects of treatments, and didn’t address risks of bias from poorly or non-randomised studies; or if they’re trying to look at prognosis and haven’t thought about drop-outs, then you can refuse to read further.)
S- Synthesis – look at heterogeneity. (When you look at how they have pulled all the information together, was it so mixed up and messy that it shouldn’t have been done the way it has? Did they assess publication or other biases? If they didn’t then send the forest plot to the composter.)
T – Transferability. This one takes a bit longer to explain. It might be a review of steroids for croup, that tells you ‘steroids work’. Hurray! But you can’t just prescribe ‘steroid’ can you? How can you transfer the review result into clinical practice? A couple of suggestions have been offered, and both have common sense behind them.
- go for the strategy of the largest trial (it has the most weight in the review).
- go for the most often used dose/drug/approach (it was shown to be feasible the most often).
So, next time you pick up a systematic review, instead of just believing it – ask some FAST questions to check.